How Does the Environment Affect thePerson?

Mark H. Bickhard

How Does the Environment Affect the Person?Mark H. Bickhard invited chapter in Children's Developmentwithin Social Contexts: Metatheoretical, Theoreticaland Methodological Issues, Erlbaum.edited by L. T. Winegar, J. Valsiner, in press.

How Does the Environment Affect the Person?

Mark H. Bickhard

Abstract

Standard conceptions of how the environment influences the person are

constrained by the dominant view of representation - and, therefore, perception,

cognition, and language - as fundamentally consisting of encodings. I argue

that this encoding view is logically incoherent. An alternative view of

representation is presented, interactivism, and shown to avoid the

incoherencies of encodingism. The interactivist model of representation

provides accounts for standard presumed encoding phenomena, and highlights

processes and forms of influence of the environment on the person that are

obscure or entirely absent from the encoding account. The multiplicity and

complexity of the processes of environmental influence acquire a theoretically

coherent organization and development from within the interactive perspective.

How Does the Environment Affect the Person?

Mark H. Bickhard

Introduction

It is generally assumed that human beings perceive and understand theworld through the senses, and that that epistemic connection with the worldoccurs via the transmission of information from the world through those sensesinto a mind. The converse perspective on this same assumption is that theenvironment influences individuals, both microgenetically and developmentally,via the information that is generated in that environment and transmitted into theminds of those individuals. I wish to contest this standard view of the nature ofepistemic contact with the world, and, therefore, also contest the correspondingstandard view of how the environment influences behavior and development.

A quick sense that there might be something wrong with both sides of thestandard view can be derived from consideration of what is usually taken to bea purely philosophical problem with purely philosophical consequences: theproblem of skepticism (Annas & Barnes, 1985; Burnyeat, 1983; Popkin, 1979;Rescher, 1980; Stroud, 1984; Wittgenstein, 1969). Briefly stated, the problem ofskepticism arises from the question: How can we possibly know that ourrepresentations of the world are correct? The only answer seems to involvechecking those representations against the world to see if they in fact match,but, by assumption, the only epistemic contact we have with the world is viathose representations themselves - any such check, therefore, is circular andprovides no epistemic ground.

Skepticism is generally relegated to philosophy, and, althoughphilosophers periodically attempt to discredit the skeptical question, no one hasin fact succeeded in solving it. The consensus, however, is that there has to besomething wrong with the skeptic's position, since it is clear that we do in facthave epistemic knowledge of the world. This presumed invalidation of thequestion, and, therefore, of the problem, is presupposed with even greater forcein psychology - not only must there be something wrong with the question thatseems to pose the problem, but it's all just philosophizing anyway and has norelevance to the business of psychology.

Unfortunately, psychology is, among other things, in the business oftrying to understand epistemic relationships between individuals and the world,and of addressing other relationships that often make strong presuppositionsconcerning the fact and the nature of such epistemic relationships. Even if weaccept the fact of such epistemic contact between the individual and the world,our models and our presuppositions commit us to particular conceptions of thenature of that epistemic contact: the simple rejection of the skeptical conclusionthat we do not have any such epistemic contact does not suffice to invalidate therelevance of the skeptical argument to psychology. In particular, if the standardpresuppositions concerning the nature of those epistemic relationships are in

2

fact vulnerable to the skeptic's argument, and if the argument is not invalid initself, then the entire body of work which involves those presuppositions isinvalidated. I will argue 1) that the skeptic's problem is one of a class of relatedproblems, all of which are valid and fundamental to the epistemologicalenterprise, 2) that contemporary approaches to epistemology - of, for example,perception, cognition, language, or sociality - are intrinsically incapable ofsolving or of dissolving these problems, and 3) that, therefore, approaches thatmake standard presuppositions concerning these epistemological issues - suchas contemporary approaches to understanding the influence of the environmenton the behavior and development of the individual - are similarly invalidated. Ithen wish to outline an approach that is not vulnerable to the general class ofproblems that includes the skeptic's problem, and to explore some of theconsequences of this approach to the general problem of the influence of theenvironment on behavior and development.

The Impossibility of Encodingism

The "transmission of information" model rests upon a general view of thenature of representation: a view of representation as consisting fundamentallyof encodings. In this view, 'information' is encoded, transmitted, decoded, andnew encodings are generated on the inferential or heuristic basis of otheralready extant encodings. In other words, information is transmitted - andprocessed and understood - in the form of encoding representations. Myrejection of this view rests on a rejection of the encodingist model ofrepresentation: if representation is not fundamentally constituted as encodings,then the transmission view cannot be sustained, and must be changed inunforeseeable ways to accommodate the non-encoding character ofrepresentation, whatever that may be. I begin, then, with a characterization ofencodingism, followed by a further elaboration of its critiques, an alternativemodel of representation, and an exploration of some consequences.

Three equivalent characterizations of encodings will be outlined:encodings as representational stand-ins; encodings as representations definedin terms of what they represent; and encodings as known correspondences withwhat they represent. The stand-in perspective on encodings is clearest andmost paradigmatic. It captures directly the character of such encodings asMorse code or computer code. The basic notion is that an encoding stands-infor some other representation, as, for example, "..." stands-in for "S" in Morsecode, or equivalently for some bit pattern in a computer. Such stand-ins changethe form of representation, and thereby allow things to be done with and torepresentations that would otherwise be impossible or difficult: "..." can be sentover a telegraph wire, while "S" cannot, and the potentialities of bit patterns incomputers are myriad. The stand-in relationship can also be defined withrespect to combinations of other representations, creating, in effect, encodingabbreviations. The critical point for my current purposes is to note thatencodings as stand-ins require that the representation(s) that are to be stood-in-for must be already present for the stand-in encoding to be definable. Stand-inencodings only change the form of representation, they do not and cannotcreate new representations (except in the sense of new combinations ofrepresentations already present).

3

The second characterization of encodings is as representationalelements defined in terms of what they represent. This is manifested instandard manners of speech such as "This thing, X say, represents (encodes)Y" where "Y" specifies what "X" is to represent. This is, in fact, the manner inwhich most encodings are introduced - they are defined as encodings byspecifying what they are to be taken as representing. This view of encodings,however, is just a different perspective on encodings as stand-ins. The definedencoding stands-in for whatever is used to specify what it represents: "X" standsin for "Y". In other words, to define "'X ' represents Y" requires that "X" bealready known, that "Y" be already known, and that what "Y" represents bealready known so that "X" can be used to represent the same thing as "Y" - sothat "X" can stand-in for "Y".

Encodings as known correspondences is still a third perspective. "X"encodes Y involves an epistemic correspondence between "X" and Y that isknown to whatever epistemic agent is able to take "X" as an encoding of Y.Such epistemic correspondences can be arbitrarily defined between any "X"and any Y, or the epistemic definition can be based on already existing factual,perhaps even lawful, relationships between "X" and Y. In order to know thecorrespondence, in order to be able to take "X" as an encoding for Y, whetherarbitrary or not, an epistemic agent has to already know both "X" and Y and,perhaps, the non-arbitrary non-epistemic (factual or lawful) relationshipbetween them. In knowing this relationship and what the relationship is with,specification of what the relationship is with must itself occur in terms of somerepresentation or another, some "Y", and, with respect to that specifyingrepresentation, "X" is again a stand-in. All three views of encodings, then, areequivalent: they are just differing perspectives on one underlying form ofepistemic relationship.

The correspondence view, however, can be particularly misleading. It isoften tempting to consider factual or lawful correspondences to constituteencodings - to constitute epistemic relationships - without explicit considerationof what the relevant epistemic agent is or how it could possibly know of thecorrespondence at all or what the correspondence is a correspondence with.Neural activity in the retina, for example, is generally in factual correspondencewith various properties of the light, and this is labelled an encoding of thoseproperties of the light. DNA base pair triples selectively correspond to particularamino acids in protein construction, and this too is labelled an encodingrelationship. Yet, there is no agent in the retina, or neural tract, that knowsanything about those light properties. Human beings and other animals havebeen seeing their environments for millions of years without knowing anythingat all about light properties per se. Nevertheless, the encoding story - thesensory transduction story - is the standard account of vision and other sensoryprocesses (Carlson, 1986).

Transduction, in its basic meaning, refers to a transformation of form ofenergy. Such a transformation will, in general, yield a factual correspondencebetween the two forms of energy and the events associated with them. Tosimply assume that this factual correspondence constitutes an epistemiccorrespondence, as in transduction models of sensory processes, is not only a

4

non-sequitur, the incoherence problem shows that it is impossible to validly fillin the argument.

Similarly, DNA base pair triples exert specific effects in a complicatedprocess of protein construction, effects that do in fact differentiate particularamino acids, and, thus, establish a factual correspondence with those aminoacids. But there is no epistemic agent involved here at all (or, if one were tocontend that there is, it is not specified what it is or how it works), and noepistemic encodings either. Note that the problem here is not with theseusages of the term "encoding" per se - semantic arbitrariness certainly allowsthat - the problem is the easy seductiveness of the invalid equivocation oftreating such non-epistemic versions of 'encodings' as constituting epistemicencodings. In the case of the sensory systems, such non-epistemic factual-correspondence transduction 'encodings' are at times even considered to beparadigmatic cases of epistemic encodings.

For a slightly different example, consider that we might speak loosely ofcertain spectral lines in sun light "encoding" various properties of and in thesun. The relevant correspondences are there, and have had to be discoveredlaboriously over centuries by astronomers and physicists, and, for one of thoseastronomers or physicists, it might even be true that those spectral lines encodeproperties of the sun, but it is clear that the encoding relationship, howevermuch it is based on lawful correspondences, is constituted in the scientist'sknowledge of those correspondences, in the epistemic correspondences, not inthe mere factual correspondences per se.

A related example is found in computer codes. The sunlight spectrallines example is based on physical law correspondences; computers involvearbitrary designer-specified correspondences. But, in both cases, thecorrespondences are known, and, therefore, the epistemic encodings exist, onlyfor the scientist, in the first case, and the designer or user in the second.Computers do not represent anything for themselves (Bickhard and Terveen,manuscript 1989).

If this general point is correct, that factual correspondences cannot inthemselves constitute encoding epistemic correspondences, then much ofpsychology is in serious trouble. Information about the world is almostuniversally assumed to enter the mind via the senses, and the senses, in turn,are with very few exceptions considered to be sensory encoding systems. Thesense in which they are considered to be encoding systems is basedfundamentally on the correspondences between neural activity andenvironmental properties. If those factual correspondences do not in fact justifythe notion of encodings, then psychology offers essentially no alternative modelof how the individual can make epistemic contact with its environment.Therefore, it equivalently offers no viable model of how that environment caneffect the individual.

I wish to argue this point at an even deeper level. It is not only that theobserved correspondences in the sensory systems do not constitute encodings,but that strict encodingism in general, in any presumed form or incarnation, is

5

logically incoherent and incapable of performing any of the standard epistemictasks that are ubiquitously assigned to it: perception, cognition, language, andso on. That is, it is not just that the relevant encodings or encoding processeshave been misidentified, it is that encodingism is a fatally flawed conception ofthe nature of representation.

The first part of the complex of arguments for this position has alreadybeen presented: skepticism. Encodings require knowledge of what they are torepresent, and, at least at a basic level, those encodings are the means bywhich we know what they are to represent. Therefore, to check their accuracy,we must check them against what they represent, which means we check themagainst themselves. This is circular, and provides no check at all.

A second, related, argument has to do not with questions ofrepresentational accuracy, but with questions of representational origin. Howdo we know which encodings to activate, to set up? They are to be set up incorrespondence with the world, but what is that correspondence to be acorrespondence with? How can we construct a copy of the world before wehave our copy of it? (Piaget, 1970) We must already represent the world beforewe are able to construct our representations of it.

A standard rejoinder to this point would be to claim that the constructionprocess is taken care of automatically in the lawful relationships established bythe sensory encodings. But, as discussed above, those lawful relationshipsestablish only factual correspondences. They do not in themselves constituteepistemic relationships; they do not provide knowledge of what thecorrespondences are with. Therefore, they are not, in fact, constructions ofencodings in any epistemic sense. Given a particular sensory neural activitypattern - a purported encoding - what is it that we are to set it up to be anencoding of?

A third consideration in the complex of arguments against encodingismderives from questioning how we are supposed to know what an encoding issupposed to represent at all. In standard cases, we know what an encoding issupposed to represent because it has been defined or specified in terms ofsome other representation - it has been defined as a representational stand-in.Those defining representations, in turn, might similarly be defined in terms ofstill other representations, and those in terms of still more basic representations,and so on. But this regress must stop at some finite level, and it is at this levelthat an incoherence is found. At the base level of representations, out of whichall other encodings are to be defined, we must have logically independentencodings in the sense of their representational content - knowledge of whatthey represent - not being provided by any other representation. If it werespecified by some other representation, then the presumed basic encodingwould not be basic. If it is not specified in terms of some other representation,then it has only itself to provide representational content, which leaves us with"'X ' represents whatever it is that 'X ' represents". This fails to provide "X" withany representational content, and, therefore, fails to constitute "X" as anencoding representation. A presumed foundational, logically independent,encoding is an impossible - logically incoherent - concept.

6

The deepest level of explanation for all of these arguments and failuresof encodingism has already been adumbrated above. Encodings are stand-ins,and stand-ins require representation to be already available to be stood-in for.Encodings are known in terms of what they represent, and knowing what theyare to represent requires prior representation of what they are to represent.Encodings are known correspondences, and known correspondences requireprior knowledge of what the correspondences are with. These are all three thesame point, just looked at from the three encoding perspectives introducedabove. The point is that encodings only change the form of already existentrepresentation. Encodings do not and cannot account for the emergence ofnovel representation out of non-representational ground (Bickhard, in press-a).Clearly such emergence occurs, both evolutionarily and developmentally, and,therefore, encodingism cannot suffice (Bickhard, 1988; Bickhard and Campbell,1989; Piaget, 1971, 1985). To presume that encodings can account for theemergence of representation, either evolutionarily or developmentally ormicrogenetically, is to presume that they can explain what they alreadypresuppose - the existence of representation. This is fundamentally circular(Bickhard, 1982). Skepticism and the origins problem and the incoherenceproblem are all versions of that basic ontological circularity in any strictencodingism.

The conclusion, of course, is that representation cannot be fundamentallycharacterized in terms of encodings. There must be some other form ofrepresentation that can solve the problem of representational emergence, andavoid the incoherent circularities. Such an alternative model of representationmight well force changes in standard notions of how persons know theirenvironments, and, conversely, how environments effect persons.

Interactivism

Any successful goal directed interactive system must manifest in itsinteractions sensitivities to the conditions in which it attempts to reach its goals.In particular, it must differentiate its activities in accordance with appropriatedifferentiations of its environments. I will argue that this interactive goal directed'sensitivity', this environmental differentiation, is the foundation of allrepresentation (Bickhard and Richie, 1983), that it is a non-encoding form ofrepresentation, and that its emergence out of non-representational phenomenais non-problematic.

An open system in interaction with its environment will proceed in thatinteraction in accordance with that environment and with the internalorganization of the system. Differing environments can yield identical, differing,or partially overlapping internal courses of activity within that systemorganization. If the system has two or more potential internal states that it mighthalt in when the interaction has ended, then those final states, say A and B, willserve to differentiate the class of potential environments into those that yieldfinal state A and those that yield final state B. That is, the set of potential finalstates of an interactive system constitute a set of implicit categories of potentialenvironments that the system can differentiate actual environments into(Bickhard and Campbell, 1989).

7

Such differentiations are not encodings. There is no stand-inrelationship. There is no knowledge of what the differentiation is adifferentiation of. There may well be a factual correspondence involved withwhatever the differentiations are in fact differentiations of, but there is noknowledge of the fact of such correspondence nor of what any suchcorrespondence is with. There is simply a system in one of several possiblefinal states. There is no representation of anything about an A-typeenvironment, for example, except for the functional sense in which it is differentfrom a B-type environment. In other words, differentiation is more primitive thanencoding.

Note that if the differentiating interactive system is in fact only passive,with no output, it can still serve a differentiating function, although itsdifferentiating power may in general be limited relative to that of the class oftruly interactive systems. For such passive differentiators, the point concerningthe possibility of factual correspondences being involved in the differentiationsholds just as for active differentiators. Such passive differentiations - perhapssimple energy transductions - that establish factual correspondences, butunknown and unrepresented correspondences, are in fact what are found in thesensory systems of organisms. The construals of perception as being based onsensory encodings, then, involve misinterpretations of the "passive functionaldifferentiations with resultant correspondences" that actually occur at sensorysurfaces as constituting known or represented correspondences, thusencodings. Physiologically identified sensory correspondences help explainwhy the functional differentiations involved can be useful to the organism - thedifferentiations are of ecologically useful properties - but their usefulness doesnot involve representation of what those differentiations are differentiations of. Itdoes not involve sensory encodings (Bickhard, in press-c; Bickhard and Richie,1983).

Interactive (or passive) differentiation, then, is a potentially useful functionthat is more primitive than encoding, yet captures at least one of the importantproperties of representation - differentiation. Furthermore, interactivedifferentiation is non-problematic in terms of how it might be accomplished - it's"simply" a functional aspect of the activity of the system - and non-problematic interms of how it might emerge - system organization need "simply" to evolve ordevelop or be designed to serve such a function.

It is not, however, in itself full representation. In particular, the sense inwhich it is more primitive than encoding is, among others, that it does notinvolve any representational content. There is no representation of what thedifferentiations are differentiations of, of what the correspondences arecorrespondences with. With encodings, the function of differentiation and thecontent of representation are inseparable: both must be present in order foreither to be present. That is one aspect of the circularities involved in a strictencodingism. In this interactive model, however, we have the emergence ofdifferentiation without representational content. Representational content is notnecessary to the function of differentiation. Accounting for the nature andemergence of representational content itself, now perhaps on the basis of non-contentful differentiation, is the next task.

8

Representational content emerges in the usefulness of thedifferentiations to the rest of the system. It emerges in the potential uses towhich those differentiations can be put. In particular, if the differentiating systemis a subsystem of a larger goal directed system, and if that larger system selects- functionally differentiates - its goal directed activities in part in terms of the priorenvironmental differentiations of the subsystem, then those subsequentpotential selections constitute predicated properties of the differentiatedcategories of environments: they constitute representational content aboutthose environments. If the goal directed system has goal G78 active, and thesubsystem has differentiated an environment of type A, then procedure P133may be selected, while if an environment of type B has been differentiated, thenprocedure P22 may be selected. Such internal selection relationshipsbetween differentiations and further activities in the service of internal goalsconstitute representational contents: "A-type environments are P133 typeenvironments" and "B-environments are P22 environments". These implicitpredications can be true or false, and are testable and tested in actuallycarrying out the interactions. The predications, thus, the representationalcontents, are implicit in and emergent from the system organization that yieldsthe initial differentiations and that selects further system activity on the basis ofthose differentiations. Those predications, and, thus, the representationalcontents, can evolve, develop, or be designed in terms of that systemorganization (Bickhard and Campbell, 1989). The principles of such systemorganization, out of which representation emerges, are not themselvesrepresentational: The problem of emergence is, in the interactive perspective,not circular, and, therefore, is not an aporia.

It is important to note that the goals involved in this model of theemergence of representational content are not themselves representational. Ifthey were, then the account would involve a circularity. All that is required forthe model is that the system have indications of success or failure functionallyavailable - without such indications, there would be no truth or falseness of theindicated interactive properties, and, therefore, no implicit predications. Forsuch functional indications of success and failure, the "goal directed" subsystemneed only be functionally organized so that it "tries again" under some internal -failure - conditions and "proceeds to other interactions" under other internal -success - conditions. With such a functional organization, the internalconditions under which the subsystem proceeds to other parts of the systemconstitute a functional goal for that subsystem - the system keeps trying untilthose internal conditions obtain, and then it goes on to other interactions. Thisnotion of goal does not itself circularly require any epistemic notions of goalsince it is a strictly internal and strictly functional explication of goal - and,therefore, of goal directedness. It is always possible, of course, that thenonrepresentational functional goal condition will itself involve subsidiaryepistemic checks on represented conditions in the environment, but any suchgoal-subsidiary representations will themselves have to be emergent in goal-subsidiary functional organizations with their own environmentaldifferentiations, implicit definitions, and interactive representational content.

Representations of (properties of) what is being differentiated, therefore,are constructible from non-representational system organization, and prior

9

representation of what is being differentiated is not required in order to constructor accomplish differentiation itself. This model, then, is not subject to theencodingist aporia of emergence, nor the incoherence of encodings, nor theproblem of origins. Still further, if the subsystem is in final state A, then thesystem is in an A type environment tautologically, and the further functionalpredications are testable in terms of further interactions that involve them. Theinteractive model involves an emergence of representation out of action in sucha way that there is an intrinsic functional relationship between representationand action, not just an interpretive relationship as with encodings. This yieldsthe consequence that representation can be tested in action, unlike encodings,for which there is an arbitrary gulf of interpretation between representation andaction. (There is no claim here that such testing yields certainty - interactiveknowledge is quite defeasible - only that such testing is possible at all.)Interactive representation is constituted as organization of potential interaction:there is no such interpretational chasm. In other words, the interactive model ofrepresentation is not subject to the argument of skepticism.

Influence in terms of Selection Pressures

In the standard encoding perspective, representation is constituted aspresumed known correspondences with that which is represented. Thisstructural correspondence model of representation invites a notion of the originsof those structural correspondences as involving a kind of passive impressionfrom the environment into the mind. This impressing of the forms of the worldinto the intellect is at least as old as Aristotle (Barnes, 1982; Kahn, 1979;Norman, 1979), and has the classic waxed slate as one of its primarymetaphors.

In the interactive view, representation is not constituted as structuralcorrespondences with what is known, but, rather, as functional interactiverelationships with what is known. It might be considered that structuralcorrespondences could be impressed into a passive mind, but the systemorganizations that constitute interactive representation could not be impressedfrom the world into the system. Interactivism requires active construction as amodel of origins.

Further, representational interactive system organization is the implicitultimate goal of such construction - it is an accomplishment that cannot beguaranteed in advance. A presumption of foreknowledge of how to constructsuccessfully representational system organization presupposes what is to beaccounted for - knowledge of the interactive properties of the world. Presciencedoes not exist. Without prescience, system constructions cannot simplyanticipate the required system organizations. System construction, then, musthave an ultimate character of blind variation constructions and selections interms of the representational interactive success of those constructions.Learning and development must have the foundational character of variationand selection constructivism (Bickhard, 1988; Bickhard and Campbell, 1989).

A related but somewhat more subtle point derives from applying thissame logic to interactions themselves, not just to constructions of interactive

10

systems. Which patterns and heuristics of interaction will succeed in achievingsuccessful interactive differentiations of the environment is, in the absence ofprescience, also something that must be tried out, perhaps multiple times, todiscover what works. What doesn't work is selected out; what does workestablishes differentiations available for potential further interaction. In fact,evolutionary epistemology - blind variation and selection - is the only possiblenon-prescient epistemology, and, in that sense, the only possible non-circularepistemology, at any level: biological evolution, historical development,individual development, or organismic interaction (Bickhard, in preparation;Campbell, 1959, 1974, 1990).

Both interactively and constructively, perceptually and developmentally,the fundamental form of influence from the environment to the person is in termsof relevant selection pressures on the interactions, and, thus, the interactiveorganizations, of the system. Sensory energy transductions, causal trajectoriesinto the nervous system, and so on, are of epistemic relevance only insofar asthey participate in such micro-genetic and developmental variation andselection constructions. Micro-genetic and developmental selection pressures,correspondingly, replace transduced encodings as the fundamental form ofenvironmental influence on the person.

Exploring the Shift from Encodings to Selection Pressures

This shift from encodings to selection pressures is not just a verbal shift,nor is it a change with limited and local implications and consequences. Acarefully developed selection pressure perspective forces at times deepchanges on understandings of 'familiar' phenomena, and new phenomenabecome apparent that are not discernable, or discernable only in dim anddistorted manner, in the encodingist perspective.

To begin, I first address the familiar fact that most of our interactions arenot unsuccessful, and most of our learning trials are not blind. We do seem tohave foreknowledge. There are two parts to the understanding of this: the firsthas to do with the nature and possibility of such foreknowledge, and the secondwith the origins of such foreknowledge.

The nature and possibility of such foreknowledge is, in both themicrogenetic and the developmental cases, grounded on the massiveredundancy of the world. By redundancy is meant the opposite of informationalindependence: a great many interactive potentialities of the world are indicatedby the outcomes of other interactions. Visual scans, for example, suffice toindicate multiple complex interactive opportunities afforded by physical surfacesand objects, by other agents, and so on. Similarly, categorization of problemsinto types can indicate the forms and strategies likely to yield solutions. Neitherthe interactions indicated by the initial visual scans, nor the problem attemptsusing strategies and heuristics indicated by the problem categorizations, will befully blind. They might, in fact, be so familiar and practiced as to seemdeterminate and certain. Much of our quotidian life involves such familiar,habitualized, and taken for granted foreknowledge of what is possible and whatwill work.

11

The origin of such knowledge, however, requires first that those relevantindications of interactive and constructive potentiality be functional in the systemorganization, and that requires that those indications have been constructed inthat system organization. That construction, in turn, may also involve priorheuristic knowledge of problem types and heuristics for solutions, but, at somepoint (whether developmental or evolutionary) those constructions cannotinvolve such foreknowledge on pain of either circularity or infinite regress. Atsome point, initial constructions must have been without any foreknowledge,must have been blind.

Similarly, even given the constructed presence of such indications in thesystem organization, they will be functionally active only if and when therelevant interaction outcomes that evoke or trigger the indication have beenreached, and those will have been reached only if the relevant interactionshave been attempted. The capacity for those relevant interactions will have tohave been already constructed, but, more important for present purposes, theattempts of those interactions, given the presence of the relevant constructions,might themselves be based on indications from prior interactions, whosepotentiality might likewise have been indicated by the outcomes of still furtherback interactions, and so on. But this regress too must terminate, and at thatpoint of termination, there must be interactions without prior indications - blindattempts at interactions. These blind levels of interactive trials will generallyinvolve microgenetic details of sensory interactions - for example, visualinteraction differentiations of distance via motion parallax, or of objects andshapes in a cluttered and unclear scene - upon which the vast complex ofunderstood redundant potentialities of our environments are perceptually andapperceptually based (Bickhard, 1980a; Bickhard and Richie, 1983).

The epistemic edge of our knowledge of the world - both of how the worldcould be in general, and how our current environment actually is - is necessarilygrounded in blind variation and selection. The great bulk of our adult world, onthe other hand, is constituted as the enormously vast and complex potentialitiesfor further interaction and construction that are indicated for us on the basis ofthose grounding interactions, that are informationally redundant with thosegrounding interactions. Conversely, we engage in much of our perceptual andlearning experiences for the sake of those redundancy based indications offurther potentiality, of what is possible and what might work (Bickhard, 1980a).

This is the replacement level for the simple encoding model: encodingsfrom the environment are replaced by selection pressures from theenvironment, operating on foundationally blind variations of micro-genetic ordevelopmental constructions. We do not encode sensory properties of theworld from which we infer perceptual and cognitive properties of the world, fromwhich, in turn, we infer action potentialities. Instead, we engage in initially blindinteractions with the world, seeking those that fit the current selection pressures,on the basis of which the vast complexes of interactive potentialities, interactiveaffordances, that constitute the interactive world are directly indicated (Bickhardand Richie, 1983).

12

At this simple replacement level, we escape the incoherencies andaporias of encodingism, and understand the emergence and function ofrepresentation in the field of action. We also understand the necessity, as wellas the actuality, of models of active mind.

Implications and Consequences

Just as constructivism and an active mind are logically forced by theinteractive model of representation, however, so also are many additionalproperties and potentialities. One critically important example derives fromnoting that the environment can influence the person not only by imposingselection pressures, but also by blocking selection pressures. The notion ofblocking selection pressures, in turn, gives rise to a functional notion ofscaffolding.

Scaffolding. Constructions of system organization are subject toselection pressures, and survive only if they are minimally successful in theirinteractions. Constructions are in the context of already present system, and, ingeneral, make use of previously constructed organization as the basis forcurrent construction. If the amount or complexity of construction necessary, inthe current context, to achieve a successful organization relative to theoperative selection pressures is too great, then non-prescient constructions areunlikely to happen-upon the successful organization. Successful constructions,in other words, tend largely to be 'small' additions to and variations on what isalready present. If such 'small' constructions are not sufficient to satisfy somearena of selection pressures, the system may be unable to achieve success inthat arena of selection pressures.

If, however, some of those selection pressures could be blocked, thensome constructions that would be inadequate with respect to the full array ofselection pressures might become adequate with respect to the reducedpressures. Points of constructive success might emerge that wereconstructively closer to the existing capacities of the system, and, thus,potentially achievable by the constructive processes. If several suchconstructively accessible points of successful, therefore stable, constructionemerge that are constructively close, they may form a trajectory of potentialconstruction, each point of which is constructively accessible from thepreceding, that ends with an otherwise not constructively accessible systemorganization that is successful and stable with respect to a full domain ofselection pressures. That is, the blocking of some selection pressures mayallow the system to climb a trajectory of stable constructions and achieve aresultant construction, and, thus, a resultant capability, that it could not havereached without the suspension of those selection pressures. Muting orblocking selection pressures in the service of variation and selectionconstruction is functional scaffolding: the generally temporary suspension ofselection pressures scaffolds the constructions of capabilities that ultimately donot require that scaffolding (Bickhard, this volume).

This conception has much in common with the standard notion ofscaffolding in the literature as the provision of knowledge that is not otherwise

13

present in order to make possible accomplishments that are otherwise beyondthe capacity of the individual, usually a child. This provided knowledge ispostulated to be eventually internalized by the child, and thereafter no longerneeds to be provided externally. Such provision of knowledge, for example, inthe form of organization and coordination of activities - each of which is withinthe capacities of the child, but for which those coordinations are not within thecapacities of the child - can permit much to be accomplished by the child andscaffolder together that are impossible to the child alone (Bruner, 1975;Vygotsky, 1978; Wertsch, 1985; Wertsch and Stone, 1985). The process ofinternalization, however, is generally left very unclear, and, by default if not byexplicit model, is left to some form of encoding by the child. The model ofscaffolding the variation and selection processes for the child offers a non-encoding alternative to this view of internalization.

The most important consequences of this functional notion of scaffolding,however, result from its being much broader than "the provision of knowledgethat is not present in the child." The muting or blocking of selection pressurescan be accomplished by the provision of otherwise absent knowledge, but itdoes not require such knowledge.

A deep and powerful example is provided by the function of selfscaffolding. This notion makes no sense whatsoever in standard conceptions:how can a child or adult provide to him or herself knowledge that he or shedoes not already have? I will argue, in contrast, that self scaffolding is anessential class of skills that children develop and even that self scaffoldingconstitutes its own field of development.

Self Scaffolding. Scaffolding is the function of reducing the demandsof otherwise perhaps too difficult tasks. Those demands - selection pressures -can be blocked, suspended, muted, set aside, compensated, or satisfied byknowledge provided externally. Only the latter version requires knowledge ofhow to satisfy the relevant selection pressures. From the perspective of oneperson scaffolding a developmental task for another, this is already broaderthan the "provision of knowledge" model. For example, much of the scaffoldingprovided to children in language learning is not the provision of otherwiseabsent knowledge, but the sometimes multi-trial processes of makingcommunications efforts on the part of the child successful that would normallybe obscure and unsuccessful. In other words, adults often devote considerableeffort to figuring out what a child has in mind.

From the perspective of an individual, including a child, any reduction incomplexity of a problem, making use of resources that are not intrinsically partof the problem situation, breaking down into subproblems, moving to idealcases, moving to simpler cases, and so on, is an instances of self scaffolding.Each provides an instance of reducing the selection pressures on successfulconstruction in the service of ultimately satisfying the full selection pressures.

All of these functions of blocking, suspending, muting, bracketing, and soon, selection pressures must be served by activities of the scaffolder - someoneelse or the individuals themselves. These are skills. Some scaffolding and self

14

scaffolding skills will be general in application, while others will be specific toparticular problem domains. All such skills must be themselves learned anddeveloped.

Self scaffolding, then, is a critical aspect of learning to learn. It is a fieldof skills that itself undergoes development. It is a field of development that isitself critical to much, if not most, other development (Bickhard, this volume).

Learning the skills and heuristics, and learning the efficacy, of selfscaffolding is itself a potentially complicated and difficult field of development.As a field of development in its own right, the development of self scaffoldingmight itself be scaffolded by caregivers and instructors. Instruction is usually amatter of presenting selection pressure experiences that provide usefulredundancies for potential future interaction demands. Good instruction is itselfwell scaffolded, though it is questionable how often this is to be found in ageneral context that views memory and cognition as passive recordings of"information". Even better teaching involves a scaffolding of the self scaffoldingof the students. This certainly occurs, but it is rare, and its grounds are intuitiveand without theoretical base.

The notion of functional scaffolding yields a broader perspective onprocesses of learning and development, and the influence of the environment ingeneral, and caregivers and instructors in particular, on that development.Scaffolding can take the form of the temporary provision of otherwise absentknowledge, but the blocking of selection pressures can be accomplished inother ways as well. Instruction in general must involve some sort of scaffolding -though the approximative processes of variation and selection constructions arenot always well served by the breaking of problems down into subproblemswhich are then taught piecemeal and in isolation. We certainly don't learn towalk by practicing submovements to perfection and then combining them.Instructional scaffolding usually leaves much to be desired.

The necessity of scaffolding and self scaffolding and the importance ofthe scaffolding of self scaffolding themselves serve as selection pressures onmethods of instruction and training and acculturation. Forms of experience thatdo not in some minimal way satisfy these selection pressures will not succeedin yielding the learning and development and acculturation of the child that isnecessary and desired - they will not survive. Conversely, properties ofinstruction, training, and socialization that in standard views may not besatisfactorily explainable may make quite good sense if understood within theperspective of these selection pressures.

Functional scaffolding not only yields enhanced notions of the demandsand possibilities of instruction, it has implications and consequences in otherdirections as well. For example, child play in part consists of the mutualscaffolding of social interaction between the children involved. Elsewhere Iargue that attachment and identification both involve important aspects of selfscaffolding on the part of the child (Bickhard, this volume). Many institutions,such as school systems or half-way houses, exist to serve scaffolding functions.Childhood in the broadest sense is a biological and social scaffolding. Later I

15

will suggest a fundamentally important role for the creation of "permanent"scaffolds - aids to interaction that are, once constructed, permanently available.

Developmental Constraints.

Selection pressures are the proximal form of influence on developmentand interaction. Not all properties of development and the person, however,can be explained in terms of selection pressures per se: there are propertiesand patterns of the selection pressures that visit themselves on developmentand the person via those selection pressures, and that can be explained only interms of explaining those properties and patterns. One form of such meta-pressure is that of developmental constraints.

Selection pressures can arise from many diverse origins. They can beinherent in an environment, deliberately imposed as in instruction, or emergentin the nature of a task, as when the tasks of instruction and acculturation arethemselves subject to the selection pressures of meeting the demands ofvariation and selection constructive processes - subject to the necessities ofminimal forms of scaffolding and the scaffolding of self scaffolding.

The most directly intuitive form of selection pressure is a directly imposedtask from an environment. Direct selection pressure is not the only kind ofconstraint on development, however, and several others are at least asimportant and much more subtle. One form of such subtlety is patterningsamong selection pressures themselves. Such patterns impose constraints onthe manner, the form, and the sequence in which constructions which satisfy theselection pressures can take place (Campbell and Bickhard, manuscript 1989).That is, they impose constraints not only on what is ultimately constructed, buton various aspects of the construction process themselves - such as sequenceof subsidiary constructions. In this sense, they are meta-selection pressures.

One form of such constraint on development results from dependenciesamong the basic natures of tasks. For example, algebra must be acquiredbefore calculus can be understood; minimal weapon handling must beaccomplished before minimal hunting can be acquired - though muchdevelopment in such cases will be reciprocal and in tandem. Conversely,errors can be made in assumptions concerning such dependencies: arithmeticcan be taught based on set theory, and calculus can be taught in terms of limittheory, but in neither case is this a necessary dependency.

Another source of constraint on constructions derives from the processesof construction themselves. The prime example here is that the variation andselection character of developmental construction itself imposes a constraint:the necessity of adequate scaffolding for development to occur. The necessary"nearbyness" of successful constructions can also yield a modularity of ultimateconstructive form, with earlier constructions forming units, modules, for laterconstructions. Such modularity is not a logically necessary constraint, however,in that earlier constructions can also be (and often are) modified by laterconstructions, not just combined by later constructions. The modularityconstraint, therefore, is a constraint tendency that is subject to exceptions.

16

A related constraint tendency of constructive development derives fromthe nature of what is being constructed. If a functionally isolated subroutinehierarchy is being constructed, for example, then lower levels of the hierarchymust be functional before higher levels can be functional. Since selectionpressures operate with respect to such functionality, lower levels must ingeneral be constructed before higher levels. This too, however, is a constrainttendency, not a logically necessary constraint.

There are at least two senses in which this tendency is not necessary.First, there is in general no logically necessary reason why any particularselection pressure must be met with a subroutine hierarchy system organizationat all, and, even if so, the particulars - the number of levels, the nature of thelevels, the principles of differentiation of the levels, and so on - of the hierarchyare still not logically determinate (Campbell and Bickhard, 1986). Second, agiven system organization is not constructed in context-free isolation, and thecontext of developmental construction can have powerful consequences. Forexample, a given subroutine may be non-functional in a particular subroutinehierarchy until the lower levels of the hierarchy are available. In a contextindependent sense, then, that subroutine could not be constructed prior to thoselower levels. But that subroutine may well be also functional in some differingsystem context, and already have been constructed in that differing context,and, therefore, be available before the lower levels of this particular subroutinehierarchy. This could, in turn, have important consequences on which systemorganizations were constructively nearby to current organizations, and,therefore, on what the ultimate system organization might end up being - itmight be quite different, for example, than the subroutine hierarchy of theisolated system. Context effects can influence constructive trajectories, and,therefore, ultimate system design. Clearly, other versions of such violations ofsimple hierarchy are possible.

Levels of Knowing. One crucial constraint on development derivesfrom the nature of the knowing or representing relationship itself. An interactivesystem can represent properties of its environment. That system itself, however,will instantiate properties which might in turn be worth representing. Thoseproperties can be interacted with, and thereby represented, from a second ordersystem interacting with the first in the same sense in which the first interacts withthe environment. The second level, of course, will also instantiate properties,which can be interactively represented from a third level, which can beinteracted with from a fourth level, and so on. The interactive model ofrepresentation logically forces an unbounded hierarchy of potential levels ofknowing and representation (Bickhard, 1973, 1978, 1980a, 1980b; Campbelland Bickhard, 1986).

Since the interactive model also logically forces a variation and selectionconstructivism, together the two characteristics force a stage sequence ondevelopment: no level of knowing can be constructed without there alreadybeing interactive organization at the level below with which it can interact.Therefore, the hierarchy of levels of knowing must be ascended, if at all, in strictsequence.

17

Note that this hierarchy, and, therefore, this developmental stagesequence, is intrinsic in the nature of knowing and representation ascharacteristics of goal directed interactive systems. It is not dependent on anyparticulars of system organization, such as subroutine hierarchies, and,therefore, it is a logically forced constraint on development. If the interactivemodel of representation is correct, then this developmental stage constraint isforced by that ontology of the nature of representation. It is (onto)logicallynecessary, and admits of no exceptions.

Note also that any subroutine hierarchies that do occur in theconstructions of the system occur within particular knowing levels. Subroutinelevels are levels of control; they are not levels of representational aboutness.An indefinite number of subroutine levels can occur within a single knowinglevel - or, in principle, though not likely in practice, none. Models assuming thatsubroutine levels, or, at least, numerologically special subroutine levels, canserve as knowing levels are common, but mistaken.

The knowing level constraint on development is not itself an influence ofthe environment on the person. It is a constraint that arises out of the nature ofknowing per se. Any environmental selection pressures, however, whosesatisfaction requires ascent through the knowing levels impose the necessity ofmeeting this constraint as one aspect of satisfaction of those selectionpressures. Since the stage constraint applies to any domain of development,and since many important representations and understandings require thereflective knowings of higher levels, the stage constraint, in at least its lowerstages, will be widely imposed.

Developments that arguably require second or higher knowing levelsinclude Piagetian conservations, notions of necessity, higher forms of roletaking, the self and identity, syllogistic and other forms of logical inference, andvalues (Campbell and Bickhard, 1986). Selection pressures and constructivepotentialities in all these domains must satisfy the knowing levels constraint.

Critical Principles. The hierarchy of knowing levels together with thevariation and selection constructivism, both forced by the underlying ontology ofinteractive representation, jointly give rise to still another possibility - one that Iargue is of far reaching consequence and fundamental importance. The basicnotion is that the individual can construct not only satisfiers of selectionpressures, but, in constructing heuristics for further construction, internalinteractive representations of selection pressures can be constructed. Theseserve the function of permitting internal trials of new constructions without thetime and risk and other cost of real world trials.

Still further, such internal selection pressures will themselves constitute adomain of development. Such principles of selection will be subject todevelopment, modification, and rejection on the basis of further interaction.

Now taking the knowing levels into account, it follows that internalselection principles are also subject to the stage hierarchy. In this hierarchy,higher such principles are about lower selection principles, in a

18

representational sense. Higher principles may affirm lower principles - forexample, deepening their ground - or they may infirm lower principles - forexample, restricting their scope. It is possible, in fact, to do both simultaneously.

The selection principles form a subhierarchy within the general knowinglevels hierarchy. They serve as the internal sources of selection, of criticism, ofpotential new constructions. They constitute the individual's knowledge of whatsorts of constructions are in error, of what kinds of errors are possible and to beavoided. When expressed, they are the principles by which thought and actionare criticized. For this reason, I call them critical principles (Bickhard, in press-b,in preparation).

Critical principles form the internal aspect of selection. Critical principlescan themselves be criticized, and, thus, they ascend the knowing levelshierarchy. Critical principles constitute knowledge of potential error, inwhatever domain. Critical principles frame current functioning of the individual,and constrain future development.

Critical principles themselves constitute an important domain ofdevelopment, subject to its own constructive, scaffolding, self scaffolding,instructional, and scaffolding of self scaffolding functions. In one sense, theyconstitute an internal scaffolding of the task of development - a muting ofexternal selection pressures via already knowing, defeasibly, what errors toavoid.

I have argued that the development of critical principles constitutes thedevelopment of rationality for that individual with respect to whatever domain(s)those critical principles apply (Bickhard, in press-b, in preparation). Apotentially rational domain, then, is any in which there are learnable sorts oferrors. An actual rational domain is any in which some knowledge of thosesorts of errors is functional. The emergence of rationality is an intrinsictendency of the variation and selection and knowing levels character ofinteractive thought.

Critical principles do not necessarily function in the same way as we areaccustomed to by the dominant encoding models of thoughts as encodedbeliefs. For example, some critical principles can function quite well withoutthere ever being rational grounds for believing them of anything. An example ofthat would be truth in science. The history of science, in which establishedtheories held for centuries have been overthrown and shown to be false, giveslittle ground for rational belief that any particular theory is in fact true.Nevertheless, it is possible to establish rational grounds for concluding thatparticular theories are false, and at times even in what sense or senses they arefalse, and this is itself an advance of science. Truth functions as a criticalprinciple, a principle of selection, even when it is not rational for it to function asa belief (Bickhard, in preparation).

I mention this example to show that the manner in which the environmentinfluences the development of rationality in this sense - and this would include,for example, logic, mathematics, chess playing, physics, farming, hunter-

19

gathering, and the decision making of nomadic life styles - is not and cannot bewell understood from a belief perspective. A dominant encoding inspired beliefperspective on rationality obscures the special character and properties of suchphenomena as critical principles, and, thereby, obscures the manner in whichrationality is taught and scaffolded and so on (or not) by cultures andinstitutions. There are other sorts of phenomena obscured as well - forexample, a ubiquitous 'implicitness' in interaction and development (seeBickhard and Christopher, manuscript, 1989, for a discussion of implicitness inpersonality).

As just one example of the kind of difference this can make, consider thefollowing. If the critical principle model of rationality is correct, then knowledge1) of what sorts of errors can be made, and 2) of how to try to avoid them will beamong the most important to teach and scaffold. Yet this is not, in contemporaryWestern culture, a focus of education. Further, a primary source of such"negative" knowledge of kinds of errors is the history of a rational field - theerrors that were actually made and the subsequent discovery of them to beerrors. Nevertheless, a historical approach to educational domains is rare. Thepossibility of such a critical-principle rationality function for oral histories,literatures, and mythologies is underappreciated.

Critical principles frame rational thought. Knowledge of how toconstructively attempt to satisfy them constitutes most of rational ability. Makinguse of such aids in learning and development constitutes yet another exampleof scaffolding. These scaffolds, however - although subject to criticism andchange themselves, and although no longer directly functional for aconstruction that has succeeded in satisfying them - nevertheless are not setaside. They are not temporary, but themselves constitute the developmentalfields of rationality.

Critical principles are intrinsically historical in the sense that they are selfapplicable, and intrinsically develop from their own affirmations andinfirmations. For any developed domain, the most important arena of thathistory is that of the society and culture in which the individual is being formed,because no isolated individual can capture the time and the scope available inthose histories. Similarly, the development of critical principles is largely socialin nature, because the most important arena of criticism of thought and criticalprinciples is itself social. The selection pressures of other ideas andreasonings, critical argument, are maximal pressures for and encouragementsof the development of rationality. This implication of the theory is consistent withthe cultures of the most important flowerings of thought and creativity: e.g.,ancient Greece (Annas, 1986), post-Chou China (Reischauer and Fairbank,1960; Schwartz, 1985), the Renaissance (Breisach, 1973; Moote, 1970; New,1969), turn of the century Vienna (Janik and Toulmin, 1973; Schorske, 1980;Smith, 1987), and so on.

Constitutive Influences

Selection pressures determine success and failure, acceptability andunacceptability, of interactions and constructions. Constraints impose

20

organization and pattern on the satisfiers of those pressures and on theirconstructions. A converse perspective on the satisfaction of selection pressuresemphasizes that the construction of a satisfier of selection pressures is theconstruction of something new in the system. Commonly, this will be a newparticular of a general kind that already exists - e.g., a new problem solvingheuristic. But in some cases this will result in the emergence of fundamentallynew ontologies, with novel properties. The construction of critical principles isone example.

There is one special case in which the constructions satisfying relevantselection pressures and constraints result in a very special ontologicalemergence - the social person. Understanding the nature and emergence ofpersonhood is fundamental to the theoretical understanding of all cultures andsocieties, and of the persons in them.

Personhood exemplifies an ontology that not only owes its constructiveorigin to relevant selection pressures, but also owes its functional nature andcontinued functional existence to the sources of those same selectionpressures. Social realities impose pressures to be able to interact with them,and becoming able to interact with them is becoming a social person. Theinfluence relationships between society and persons, then, include not onlyselection pressures, scaffoldings, and constraints, and the emergence of newontologies, they also include the determination of what a person is in thatculture and the possibility and the resources for becoming and being such aperson. The influence of society and culture on personhood is constitutive.

Social Reality and Social Development. The existence of socialand cultural realities poses difficult theoretical problems. Fundamentally, whatis the nature, the ontology of such social realities, and how do they influencepersons? As adumbrated, this involves a form of influence not only on whatpersons do and think, but a constitutive influence on what persons are.

Note first that accomplishment of social interactions is a task of potentiallyenormous complexity. Interactions involving other agents pose difficultepistemological problems. How is anyone to know what sorts of interactionsare potential and what the constraints of such interactions are? Theepistemically purest form of this question poses it in terms of the social "other"being an unknown agent. The quickest answer is, of course, that we know onthe basis of multitudinous indications of body, dress, gesture, location, andlanguage - on the basis of the redundancies available, just like any otherapperception of a situation. But that only moves the epistemic question backone step: how do we know what all these things mean, how do we know therelevant redundancies? It also highlights the interesting question of why weneed them.

The first critical point here is that the redundancies available with respectto other agents are variable - in terms of agents' intentions, moods, beliefs,expectations, and so on - in ways that non-agent redundancies are not. Thesecond critical point is that the redundancies of dress, gesture, language, andso on, are themselves constructed redundancies, unlike those for rocks, trees,

21

and furniture. They are not intrinsic to being an agent per se. There is areflexivity involved in this variable constructiveness that is both ontologicallyand epistemologically unique to situations with other agents: social situations.

I will not develop the arguments here. But the general conclusion that isdrawn is that any social encounter constitutes a problem of cooperation, ofcoordination (Bickhard, 1980a). A problem of what sort of situation "this" issupposed to be. It turns out that the problem has a complex intrinsicallyreflexive organization - for example, the nature of the situation for one party to itdepends on how the other party takes it, but how the other party construes thesituation depends on how he or she takes the first party to characterize it, andso on. Escaping such occasionally dizzying iterations requires powerful tools.

In effect, it requires a scaffolding of the situation for the parties involved inorder to achieve successful interactions in the face of the unboundeduncertainties of the sorts of iterations mentioned above. I call such scaffoldssituation conventions - conventions concerning what the situation is - and arguethat they form the emergence of social realities out of individual level realities(Bickhard, 1980a).

Situation conventions are basically commonalities of assumption amongthe participants to a situation about what sort of an interactive situation theparticipants are jointly constituting. That commonality of assumption about whatthe social situation is is what constitutes the social situation. Situationconventions collectively constitute social reality, and, as scaffolds, arepermanent. They enable and they constrain social interaction by charting, and,at times, exploiting, the unknownness of raw encounter with another agent.Their permanence is a consequence of the unboundedness of theepistemologically raw social encounter in that that unbounded problem cannotbe solved in itself - only the scaffolded social situation is interactively tractable.

Situation conventions are related to contemporary notions of convention,the symbolic interactionist's definition of the situation, Goffman's frames, theethnomethodologist's reflexivities, and so on (Bickhard, 1980a). The theoreticalnotion of situation convention, however, is broader than those.

Situation conventions are scaffolds of social interaction. They permit andguide interaction that would otherwise be pragmatically impossible. They also,however, are a level of reality, of ontology, in themselves. They constitute theorganization of society and culture and thought within which the individualmoves, acts, and develops. Because the individual is his or her interacting withthe world - his or her resources for such interactions - and because much of thatworld is constituted at this level of social ontology, much of what constitutes theperson will be afforded and provided and constrained by the possibilitiesavailable in the situation conventions of the society in which the individualdevelops. These include the skills (Rogoff, 1989), the notions of roles andinstitutions (Berger and Luckmann, 1966), the values, the ideologies andlegitimations and myths (Berger and Luckmann, 1966), the social andhistorically developed emotions (Harré, 1986), the various styles of relating, andthe particular families and relationships that participate in the constitution of the

22

person. A person is not a context independent agent who merely acts in socialand cultural contexts, but is rather constituted in terms of his or her potentialitiesfor acting and interacting, and those potentialities of interaction exist largely interms of the social and cultural realities in which they participate (Geertz, 1973,1983; Harré, 1984; Shweder and LeVine, 1984).

One of the most powerful influences of the environment on the person,then, is that it is the cultural environment that determines what a person in thatculture is. This constitutive influence exists in two senses: constructive andinteractive. Constructively, learning to engage in the simpler social interactionsof childhood provides the scaffolded resources for the eventual construction ofthe adult social person. Interactively, the person is being social insofar as he orshe is interacting with or within those social realities. A social person, then, isconstituted as an entity that is capable of interacting with, thus participating in -thus jointly constituting - social realities: the constitution of a social person is theability to constitute social realities.

What social realities are, and, therefore, what persons are, will vary withculture and history. Personhood, in being a socially constituted constructiveemergent, is itself a social and historical ontology. What persons are, and whatpersons can be, evolve and change socially and historically.

Still another form of emergent constraint arises here. Societies mustsolve the problem of generations: they must nurture the constructive emergenceof new persons as constitutive participants of themselves.. They must generatenew members in order to continue. There is an inherent dynamic in thisprocess of socialization that plays its own function in history, and an inherent setof constraints in that dynamic. In particular, societies must be constituted inways that are scaffoldable and scaffolded for their new generations. The task ofbecoming an adult member of a society is paradigmatically one that is notsolvable with one massive construction. Societies must have inherent in themboth the possibilities and the actualities of decomposition into simpler modules,trajectories of successive approximations to satisfactory competence, toleranceof initially incompetent trials, and so on. Conversely, societies cannot deviatefrom such properties to too great a degree. For example, they cannot requireexcessive unlearning and relearning in moving from childhood to adulthood, orimpose grossly inconsistent characterizations of the world and its rationales(Berger and Luckmann, 1966).

Language. The situation conventions of society constitute a majordomain of interaction. Situation conventions are themselves participants in thehierarchy of reflections of the knowing levels. In particular, it is possible forsome situation conventions to provide possibilities for interacting with, ofoperating upon, other situation conventions. All human societies, in fact,provide powerful institutionalized conventions for such social realityinteractions. Elsewhere, I argue that such institutionalized conventions foroperating on social realities constitute language (Bickhard, 1980a, 1987;Bickhard and Campbell, manuscript 1989).

23

Language, then, constitutes the primary means for introducing, creating,and transforming social realities. In terms of the opportunities for andconstraints on further language interactions - for further conversation - languagealso constitutes much of most of the social realities upon which it interacts.Insofar as the person is largely social, it follows that the person is largelyconstituted at a level of language. Language not only influences the person inparticular interactions, language developmentally influences the person byconstituting the person.

This is similar to the position of the hermeneuticists (Gadamer, 1975;Heidegger, 1962; Warnke, 1987). It differs, however, in that the ground for theemergence of such social and linguistic realities, and of the persons constitutingand partially constituted by them, remains itself non-linguistic. The limits, thehorizons, of the world are not bounded solely by language. The groundremains that of a general agent interacting with an environment that is notnecessarily all linguistically, socially, constituted. Language is largelyconstitutive, but is not fully constitutive, of either a person or that person's world;language and persons are emergent out of non-linguistic reality (Bickhard,1987; Campbell and Bickhard, 1986).

Language, then, provides and constrains much of the opportunities forbecoming a person. Language is its own domain of interaction and ofdevelopment, intimately and constitutively related to social development, and,therefore, to person development. Conversely, in a manner that is nowbecoming familiar, language must be scaffoldable and scaffolded. It mustsatisfy the constraints of being able to create a new generation of constitutiveand constituted participants in that language.

Note that the microgenesis of language cannot be the standard view ofthe encoding of mental contents, their emission and transmission to anaudience, and their decoding by that audience. Utterances, like any otherobjects of perception, cannot be encoded into the nervous system, there to bedecoded into meanings and understandings. Utterances are interactions withsocial realities, and language understanding is a matter of interactiveperception and apperception of the indications of utterances for those socialrealities (Bickhard, 1980a).

Utterances, then - no more than anything else - do not influence theperson via the transmission of encoded information. For many considerations,this encoding transmission model seems to be adequate, just as the phlogistontheory of fire is consistent with many chemical interactions, but it cannot beultimately correct. The incoherence of encodingism makes an encodingistmodel of language impossible.

The sense in which utterances are understood not by a process ofdecoding, but by an intrinsically variation and selection process of interactionsand apperceptions, just as for other "perceptual" processes, is often not directlyevident. Much adult language understanding is of the well practiced andhabituated variety that needs only an initial satisfactory interactive trial to beable to complete the interaction - it appears to be algorithmic. But the

24

underlying variation and selection realities show up whenever languageunderstanding is difficult. Such difficulty can be manifested, for example, inattempting a garden path or ambiguous sentence, or a difficult text,understanding a person in psychotherapy, or learning a language as a child. Inall such cases, understanding proceeds, not algorithmically, but with trials anderrors, and shifts of considerations among various aspects of the text,attempting to find an interpretation, an understanding, that satisfies the selectionpressures of the words, the sentences, the text, the persons, and the context.This process has come to be called the hermeneutic circle (Heidegger, 1962;Gadamer, 1975).

The standard encoding view utterly obscures this process.Correspondingly, it makes it difficult to understand the processes of languageacquisition in children and the processes of instruction, training, modeling,socialization, and acculturation of children, among other things (e.g., it makes itessentially impossible to understand the development of personality andpsychopathology, and the processes of psychotherapy, Bickhard, 1989a,1989b; Bickhard and Christopher, manuscript 1989). For example, the"hypothesis testing" of children in language acquisition is intrinsic, not ad hoc,from the variation and selection view. Similarly, children's creative usages oflanguage are not just errors or mysteries, but are again intrinsically to beexpected when viewing utterances as operators - Wittgenstein's tools (Bickhard,1987) - and considering what a child could be expected to do when attemptingan utterance-operator for which he or she does not have a standard vocabulary,a standard tool - e.g., "I'm souping." or "Will you chocolate my milk?" or "I canbutton it." [turn on a pocket calculator] (Clark, 1982). It's the social operatoryversion of using a wrench as a hammer because there is no hammer available -childrens' linguistic tool boxes are limited. Further, the success of such 'errors'in communicating is likewise a mystery from the encoding perspective, as ismost metaphor and other creative language, while, in considering languageunderstanding as a hermeneutic context-embedded constructive process, thein-principle mystery of creative language disappears, even though thecomplexity remains: The words and sentences are used to evoke atransformation of the social situation, of the commonality of understanding, andto constrain - to differentiate and discriminate, to impose selection pressures on- what that transformation of socially common understanding might be. Creativelanguage also provides examples of microgenetic scaffolding: a sentence,perhaps of a poem, for example, that might be impossible to understand inisolation can in context have appropriate "understanding sets" already availablefrom earlier parts of the context such that the interpretation required tounderstand in context is no longer constructively unattainable.

The potentialities for language activity constitute much of social realityand much of human reality. The encoding view captures at best only the mostordinary and instrumental, non-constitutive, aspects of that. Language is thesource and the means of the most important influences of the environment onthe person. It will not be understood within dominant contemporaryperspectives. Among others, neither the operative character of utterances northe hermeneutic character of language understanding makes sense from withinan encoding perspective.

25

Conclusion

The primary attempt of this chapter has been to evoke a sense of dis-ease concerning standard perspectives on representation, and, therefore, onperception, cognition, and language as well. Correspondingly, standardperspectives on the nature and manner of the influence of the environment onthe person are brought into question.

An alternative model of the nature of representation is presented, andshown to avoid the logical inadequacies of encodingism. This interactivealternative not only provides coherent accounts of phenomena that areaddressed by, but, nevertheless, constitute aporia for, encodingism, it alsoyields new phenomena and new models that cannot be understood, or at bestcan be given ad hoc approaches, within the encodingist perspective. Theseinclude the imposition of selection pressures, instruction as a deliberateexploitation of redundancies, functional scaffolding, self scaffolding,developmental constraints, levels of knowing, rationality and critical principles,situation conventions, constitutive influences on personhood, and the operativeand hermeneutic character of language.

Presupposition of the encodingist perspective seduces the investigationof the influence of the environment on the person to focus on the imagineddetails of the encoding and decoding processes, and of the supposed contentsthat are being encoded and decoded. I hope that, at a minimum, this chapterhas indicated that the encodingist presupposition cannot be simply taken forgranted, and that the processes of environmental influence, not just thecontents, are of fascinating and fundamentally important complexity. Stillfurther, I hope to have offered an approach and a framework for a coherenttheoretical understanding of the influence of the environment on the person.

26

References

Annas, J. (1986). "Classical Greek Philosophy." In J. Boardman, J. Griffin, O. Murray (Eds.) TheOxford History of the Classical World. Oxford: Oxford U. Press.

Annas, J., Barnes, J. (1985). The Modes of Scepticism. Cambridge University Press.

Barnes, J. (1982). Aristotle . Oxford University Press.

Berger, P. L., Luckmann, T. (1966). The Social Construction of Reality. New York: Doubleday.

Bickhard, M. H. (1973). A Model of Developmental and Psychological Processes. Ph. D.Dissertation, University of Chicago.

Bickhard, M. H. (1978). The Nature of Developmental Stages. Human Development, 21 , 217-233.

Bickhard, M. H. (1980a). Cognition, Convention,and Communication. New York: Praeger.

Bickhard, M. H. (1980b). A Model of Developmental and Psychological Processes. GeneticPsychology Monographs, 102, 61-116.

Bickhard, M. H. (1982). Automata Theory, Artificial Intelligence, and Genetic Epistemology.Revue Internationale de Philosophie, 36, 549-566.

Bickhard, M. H. (1987). The Social Nature of the Functional Nature of Language. In M. Hickmann(Ed.) Social and Functional Approaches to Language and Thought (pp. 39-65). NewYork: Academic.

Bickhard, M. H. (1988). Piaget on Variation and Selection Models: Structuralism, logicalnecessity, and interactivism Human Development, 31, 274-312.

Bickhard, M. H. (1989a). The Nature of Psychopathology. In L. Simek-Downing (Ed.)International Psychotherapy: Theories, Research, and Cross-Cultural Implications. NewYork: Praeger.

Bickhard, M. H. (1989b). Ethical Psychotherapy and Psychotherapy as Ethics. New Ideas inPsychology, 7(2), 159-164.

Bickhard, M. H. (in preparation). From Epistemology to Rationality.

Bickhard, M. H. (in press-a). The Import of Fodor's Anticonstructivist Arguments. In L. Steffe(Ed.) Epistemological Foundations of Mathematical Experience. New York: Springer-Verlag.

Bickhard, M. H. (in press-b). A Pre-Logical Model of Rationality. In L. Steffe (Ed.) EpistemologicalFoundations of Mathematical Experience. New York: Springer-Verlag.

Bickhard, M. H. (in press-c). Cognitive Representation in the Brain. In Encyclopedia of HumanBiology. Academic Press.

Bickhard, M. H. (this volume). Scaffolding and Self Scaffolding: Central aspects of development.

Bickhard, M. H., Campbell, R. L. (1989). Interactivism and Genetic Epistemology. Archives dePsychologie, 57(221), 99-121.

27

Bickhard, M. H., Campbell, R. L. (manuscript, 1989). Some Foundational Questions ConcerningLanguage Studies: With a Focus on Categorial Grammars and Model Theoretic PossibleWorlds Semantics.

Bickhard, M. H., Christopher, J. C. (manuscript, 1989) The Influence of Early Experience onPersonality Development.

Bickhard, M. H., Richie, D. M. (1983). On the Nature of Representation: A case study of James J.Gibson's theory of perception. New York: Praeger.

Bickhard, M. H., Terveen, L. (manuscript, 1989). The Impasse of Artificial Intelligence andCognitive Science.

Breisach, E. (1973). Renaissance Europe. New York: Macmillan.

Bruner, J. S. (1975) The Ontogenesis of Speech Acts. Journal of Child Language, 2, 1-19.

Burnyeat, M. (1983). The Skeptical Tradition. Berkeley: University of California Press.

Campbell, D. T. (1959). Methodological Suggestions from a Comparative Psychology ofKnowledge Processes. Inquiry, 2, 152-182.

Campbell, D. T. (1974). Evolutionary Epistemology. In P. A. Schilpp (Ed.) The Philosophy ofKarl Popper. LaSalle, IL: Open Court.

Campbell, D. T. (1990). Epistemological Roles for Selection Theory. In N. Rescher (Ed.)Evolution, Cognition, and Realism . Lanham, MD: University Press.

Campbell, R. L., Bickhard, M. H. (1986). Knowing Levels and Developmental Stages. Basel:Karger.

Campbell, R. L., Bickhard, M. H. (manuscript, 1989). Types of Constraints on Development: AnInteractivist Approach.

Carlson, N. R. (1986). Physiology of Behavior. Boston: Allyn and Bacon.

Clark, E. V. (1982). The Young Word Maker: A case study of innovation in the child's lexicon. InE. Wanner, L. Gleitman (Eds.) Language Acquisition. Cambridge University Press, 390-428.

Gadamer, Hans-Georg (1975). Truth and Method. New York: Continuum.

Geertz, C. (1973). The Interpretation of Cultures . New York: Basic Books.

Geertz, C. (1983). Local Knowledge. New York: Basic Books.

Harré, R. (1984). Personal Being. Harvard University Press.

Harré, R. (1986). The Social Construction of Emotions. Oxford: Basil Blackwell.

Heidegger, M. (1962). Being and Time. New York: Harper & Row.

Janik, A., Toulmin, S. (1973). Wittgenstein's Vienna. New York: Simon and Schuster.

Kahn, C. H. (1979). Sensation and Consciousness in Aristotle's Psychology. In J. Barnes, M.Schofield, R. Sorabji (Eds.) Articles on Aristotle 4: Psychology and Aesthetics. London:Duckworth.

28

Moote, A. L. (1970). The Seventeenth Century. Lexington, MA: Heath.

New, J. F. H. (1969). Renaissance and Reformation. New York: Wiley.

Norman, R. (1979). Aristotle's Philosopher-God. In J. Barnes, M. Schofield, R. Sorabji (Eds.)Articles on Aristotle 4: Psychology and Aesthetics. London: Duckworth.

Piaget, J. (1970). Genetic Epistemology. New York: Columbia.

Piaget, J. (1971). Biology and Knowledge. Chicago: University of Chicago Press.

Piaget, J. (1985). The Equilibration of Cognitive Structures: The Central Problem of IntellectualDevelopment. Chicago: University of Chicago Press. Translated: T. Brown and K.Thampy. (Originally published 1975).

Popkin, R. H. (1979). The History of Scepticism. Berkeley: University of California Press.

Reischauer, E. O., Fairbank, J. K. (1960). East Asia: The Great Tradition. Boston: HoughtonMifflin.

Rescher, N. (1980). Scepticism. Totowa, NJ: Rowman and Littlefield.

Rogoff, B. (1989). "The Social Context of Cognitive Development: Building on Vygotsky andPiaget." Plenary Address at the 1989 meeting of the Jean Piaget Society, June 2,Philadelphia.

Schorske, C. E. (1980). Fin-de-Siecle Vienna. New York: Knopf.

Schwartz, B. I. (1985). The World of Thought in Ancient China. Harvard University Press.

Shweder, R. A., LeVine, R. A. (1984). Culture Theory . Cambridge University Press.

Smith, B. (1987). Austrian Origins of Logical Positivism. In B. Gower (Ed.) Logical Positivism inPerspective. Totowa, NJ: Barnes and Noble.

Stroud, B. (1984). The Significance of Philosophical Scepticism. London: Oxford UniversityPress.

Vygotsky, L. S. (1978). Mind in Society. Cambridge: Harvard University Press.

Warnke, G. (1987). Gadamer: Hermeneutics, Tradition, and Reason. Stanford University Press.

Wertsch, J. V. (1985). Vygotsky and the Social Formation of Mind. Cambridge: HarvardUniversity Press.

Wertsch, J. V., and Stone, C. A. (1985). "The Concept of Internalization in Vygotsky's Account ofthe Genesis of Higher Mental Functions." In J. V. Wertsch (Ed.) Culture, Communication,and Cognition: Vygotskian Perspectives. New York: Cambridge.

Wittgenstein, L. (1969). On Certainty. New York: Harper & Row.